NUTRIENT RETENTION CALCULATION
Comparisons of Methods for Calculating Retentions of Nutrients in Cooked Foods
Elizabeth W. Murphy,* Patricia E. Criner,’ and Brucy C. Gray
Six types of weight changes that occur when food ture-free raw and cooked foods). Comparisons in-
is cooked by different methods are described. For cluded retentions of proximate components, min-
five of these types, true retentions of nutrients erals, and vitamins. Apparent retentions overesti-
(defined as calculations based on nutrient content mated the true retentions in nearly all instances.
of known weights of food before and after cooking) To avoid bias, true retentions should be reported
were compared with apparent retentions (defined whenever it is feasible to obtain data on weights of
as calculations based on nutrient content of mois- foods before and after cooking.
Accurate knowledge of the nutrient intake of individuals Several types of weight changes occur when food is
and groups of people requires information on the nutrient cooked. These are: (type 1) volatiles (primarily moisture)
content of cooked foods. Many dietary calculations are lost; example, vegetables cooked by steaming; (type 2)
made on the basis of foods as brought into the kitchen. moisture gained; example, rice cooked so that all of the
Factors are needed that can be applied to weights of raw water is absorbed; (type 3) solids lost but moisture gained;
foods to correct for nutrient losses or changes in prepara- example, dry legumes cooked in water which is not com-
tion. This paper presents some of the problems encoun- pletely absorbed; remaining liquid is discarded; (type 4)
tered in establishing accurate retention factors. solids and moisture both lost; example, organ meats cooked
A true retention should measure the proportion of nutri- in water; (type 5) solids and moisture lost from more than
ent remaining in the cooked food in relation to the amount one tissue; example, roasted poultry, which contains lean
of that nutrient originally present in a given weight of the muscle, skin, and sometimes depot fat; (type 6) moisture
food before cooking. Thus, the direct measure of true re- lost and fat or other solids gained; example, doughnuts and
tentions requires data on the weights of the food both be- other foods fried in deep fat.
fore and after cooking, as well as the contents of the nutri- Data from research done under the sponsorship of the
ent per gram (or other unit of weight) of raw and cooked Agricultural Research Service have provided the opportu-
food. nity to compare true retentions with apparent retentions
To provide maximum useful data, studies on retentions on the same food samples for a number of foods which
of nutrients in foods should be planned so that analyses are show changes with cooking of types 1 through 5. These data
made on comparable raw and cooked samples. For meats, compare apparent retentions (AR), calculated as follows:
fish, and poultry, anatomically matched cuts representing
% AR = [nutrient content per g of cooked food (dry basis)]/
opposite sides of the same carcass should be analyzed raw
and after being cooked. From well-mixed lots of raw foods [nutrient content per g of raw food (dry basis)] X 100
such as vegetables, legumes, and shellfish, subsamples for with true retentions (TR), calculated as follows:
cooking should be carefully drawn, and similarly chosen % T R = (nutrient content per g of cooked food X g of food
subsamples should be analyzed raw. Weights of products after cooking)/(nutrient content per g of raw
before and after cooking should be recorded, along with
weights of drippings, cooking water, or other discard. food X g of food before cooking) X 100
Weights and analyses of discard are needed if a total ac- Data for cooked foods used in the calculations of reten-
counting for all nutrients originally present in the raw food tions did not include the nutrient content of any cooking
is sought, so that solubility losses, as well as destruction, discard such as drippings. Table I gives references to the
are known. Keeping records of weights may not always be analytical methods used in obtaining the nutrient data
feasible in studies involving production-line processing, but from which retentions were calculated.
should be possible in research involving institutional and
home cooking. Unfortunately, few studies have been re- TYPE 1 CHANGE, MOISTURE LOSS ONLY
ported which were designed to provide the information just
described. Retentions were calculated for five nutrients in 4 to 27
T o circumvent problems associated with obtaining vegetables which had been cooked by steaming. Vegetables
weights, many researchers have reported apparent reten- were cooked in aluminum pans over boiling water. No salt,
tion values. The apparent retention is here defined as the fat, or other ingredients were added. The few grams of
ratio of nutrient content in the cooked food without discard water that condensed in the cooking pan during steaming
to nutrient content in the raw food, with both values ex- were included as part of the cooked sample. Information on
pressed on the moisture-free basis. The use of apparent cooking time and degree of doneness was not available.
rather than true retentions involves the assumption that Table I1 shows data comparing true retentions with ap-
solids are not lost to any practical extent with cooking. parent retentions. For these vegetables, retentions were es-
This assumption is clearly not valid for meats, which give sentially complete, and differences between the two calcu-
up both fat and protein to the drippings when cooked; it is lation procedures were not significant, as indicated by
probably not valid for vegetables, legumes, and many cere- paired “t ” tests.
al products, either. Data on true retentions for 13 nutrients in six lots of
oven-roasted peanuts have been published (Derise et al.,
1974). In addition, the present authors calculated apparent
retentions for these same samples. Peanuts were roasted in
Nutrient Data Research Center, Consumer and Food the shell in an electric oven a t 177 “C (350 O F ) for 35 min.
Economics Institute, Agricultural Research Service, U.S. The shelled kernels, including skins, of both raw and roast-
Department of Agriculture, Federal Building, Hyattsville, ed peanuts from the same lot were weighed and analyzed.
Maryland 20782. Retentions calculated from the results of the analyses are
Deceased. shown in Table 111. For peanuts, unlike steamed vegeta-
J. Agric. Food Chem., Vol. 23,No. 6,1975 1153
MURPHY, CRINER, GRAY
Table I. References to Methods of Analysis Table 111. Apparent a n d T r u e Retentions of
for Nutrients Used in Retention Calculations Nutrients in Six Samples of Oven-Roasted Peanuts
(Type l a Weight Change with Cooking)
Nutrient Reference
Apparent retention True retentionb
Proximate components AOAC (1970)
(protein, fat, a s h , crude fiber) Nutrient Mean,% C.V.’ Mean,% C.V.
Mineral elements
Ca, Cu, Fe, Mg, Mn,K,Na, Zn Perkin- Elmer Protein 100 1.4 97 1.5
(1968) Fat 99 0.9 96 1.1
P (in turkey) AOAC (1970) Ash 129 5.1 124 5 .O
P (in peanuts and legumes) Fiske and Subbarow Crude fiber 105 2.5 102 2.2
(1925) Calcium 102 1.2 98 1.6
B vitamins AOAC (1970) Copper 102 2.7 98 2.6
(thiamine, riboflavine, niacin) Iron 101 2.9 98 2.8
Chole ste r ol Tu et al. (1967) Magnesium 102 1.2 98 2.6
Ascorbic acid Freed (1966) Manganese 107 2.4 104 2.7
Carotene AOAC (1970)O Phosphorus 113 2 .o 109 2.4
Retinol Ames et al. (1954) Potassium 99 0.4 96 0.8
0 Extraction procedure 39.015 modified for analysis of liver. Sodium a4 3.6 82 3.1
Zinc 104 1.3 100 1.2
Table 11. Apparent a n d True Retentions of a Volatiles (primarily moisture) lost. * All differences between ap-
Selected Nutrients i n Steamed Vegetables parent and true retentions significant ( P = 0.05). Coefficient of
variation.
(Type la Weight Change with Cooking)
Apparent
Table IV. Apparent a n d True Retentions of
retention True retentionb
Selected Nutrients in One Sample of Brown Rice
No. of (Type 20 Weight Change with Cooking)
sam- Mean, Mean,
Nutrient ples % C.V.’ % C.V. Apparent True
retention, retention,
Ash 4 102 1.7 99 1.6 Nutrient o/o %
Calcium 4 97 7.2 94 3.8
Magnesium 19 96 .
19 96 1.2 Protein 101 105
Potassium 27 103 2.8 101 2.3 Crude fiber 115 117
Sodium 20 97 6.0 97 5.5 Potassium 93 96
Volatiles (primarily moisture) lost. Differences between ap- a Moisture gained.
parent and true retentions not significant. Coefficient of varia-
tion.
Table V gives data on apparent and true retentions for
bles, differences between apparent and true retentions nutrients in the cooked dry legumes. Differences between
were significant ( P = 0.05). In all cases, the apparent reten- the two methods of calculation were significant ( P = 0.01)
tion value was higher than the true retention value. according to paired “t” tests. For all 12 nutrients, the ap-
parent retention gave a value significantly higher than the
TYPE 2 CHANGE, MOISTURE GAIN ONLY true retention. Average differences between the two meth-
Data were available for three nutrients in one sample of ods ranged from 6 to 11 percentage points, mean difference
brown rice. The rice was rinsed once with tap water and 8%,for the different nutrients.
drained before being cooked by boiling. All cooking water
was absorbed by the rice. Time of cooking was not re- TYPE 4 CHANGE, SOLIDS AND MOISTURE BOTH LOST
ported. Retentions were calculated on nine lots each of turkey
Retentions calculated from these data on cooked brown gizzard, heart, and liver, which had been cooked by sim-
rice are shown in Table IV. In contrast to retentions for mering. The giblets were cooked in distilled water in one
type 1 foods, the apparent retentions for brown rice tended pan. Because the time of cooking required was greater for
to be lower than the true retentions. Because data were gizzards than for livers, and because cooking of all giblets
available for only one sample, it is not possible to tell was started at the same time, livers were probably over-
whether or not apparent and true retentions differed sig- cooked.
nificantly for this type 2 change. Apparent and true retentions of 19 nutrients in turkey
livers are given in Table VI. These data indicate a consider-
TYPE 3 CHANGE, SOLIDS LOST AND MOISTURE able difference between the two methods of calculation.
GAINED Apparent retentions ranged from 6 to 24 percentage points,
For the type 3 cooking change, data were available on with a mean of 14 percentage points, higher than true re-
three lots each of ten different dry legumes-Great North- tentions. All differences between calculation methods were
ern, navy, pinto, red kidney, large lima, baby lima beans, significant ( P = 0.01). Thus, apparent retentions were not
cowpeas (blackeyes), chickpeas (garbanzos), green split reliable measures of the true retentions for this food show-
peas, and lentils. The legumes were purchased in local food ing type 4 changes.
markets in Virginia, and were simmered in glass cooking For turkey gizzards and hearts, differences between cal-
pans in 885 to 1189 g of deionized water, the amount of culation methods were even greater than for livers. (Data
water varying with the kind of legume. Time of cooking are not tabulated, but available from authors on request.)
ranged from 20 min for green split peas to 140 min for Comparisons of 16 nutrients in gizzards (protein, fat, ash,
chickpeas. The ratio of weights for the cooked to the dry three B-vitamins, nine mineral elements, and cholesterol)
forms ranged from 2.1:lfor chickpeas to 2.9:l for lentils. showed a range of 6 to 22 percentage points, with a mean of
1154 J. Agric. Food Chem., Vol. 23,No. 6, 1975
NUTRIENT RETENTION CALCULATION
Table V. Apparent a n d True Retentions of Table VI. Apparent a n d T r u e Retentions of
Nutrients in 30 Samples of Boiled Mature Dry Selected Nutrients in Simmered Turkey Livers
Legumes (Type 3= Weight Change with Cooking) ( 5 p e 4a Weight Change with Cooking)
Apparent retention True retentionb Apparent True
retention retentionb
Nutrient Mean,% C.V.' Mean,% C.V. No. of
sam- Mean, Mean,
Protein 103 0.9 96 1.6 Nutrient ples 5% c.v.~ % C.V.
Fat 111 3 .a 102 3.4
Ash 86 1.9 a0 3.2 Protein 9 103 1.7 a5 3.8
Fiber 134 3.4 123 3 .O Fat 9 131 5.7 107 4.7
Calcium loa 2.2 100 2.6 Ash 9 71 3 .a 59 5.3
Copper 97 2.3 90 2.9 Thiamine 9 70 11.6 58 13.3
Iron 120 2.7 111 2.7 Ribof lavine 9 57 9.6 47 10.9
Magnesium a5 3.6 79 4.1 Niacin 9 53 a .6 43 8.1
M ang anes e 104 3.4 97 4 .O Cholesterol 9 112 4.7 92 4.6
Phosphorus 92 3.4 86 4.1 Ascorbic acid 9 36 9.6 30 9.2
Potassium 98 4.9 91 5.2 Carotene 9 109 11.2 a9 9.8
Zinc 120 2.2 112 2.9 Retinol 9 62 13.4 51 15.0
a Solids lost, but moisture gained. * All differences between ap- Calcium 9 123 6.3 102 8.2
parent and true retentions significant (P = 0.05). Coefficient of Copper a 95 5.7 78 5.1
variation. Iron 9 65 7 .a 54 10.4
Magnesium 9 66 a .7 55 9.3
13 percentage points, in differences between the two meth- Manganese 8 74 5.7 61 7.1
ods. For all 16 nutrients, differences were significant (P= Phosphorus 9 76 4.4 63 5.9
0.01).For turkey hearts, differences between the two calcu- Potassium 9 57 6.2 47 7.9
lation methods were significant ( P = 0.01) for 10 of the 16 Sodium 9 57 5 .a 47 7.O
nutrients. For the remaining six nutrients (riboflavine, nia- Zinc a loa 3 .O 89 3.6
cin, cholesterol, copper, manganese, and zinc), the number a Solids and moisture both lost. All differences between ap-
of comparisons involved was small, ranging from two to parent and true retentions significant ( P = 0.01). Coefficient of
five. Had there been a larger number of comparisons, it is variation.
likely that differences between calculation methods would
have been significant for these six nutrients also, as the dif-
ferences were large, ranging from 15 to 40 percentage tween the two calculation procedures ranged from 0 to 2,
points. Thus, for foods undergoing these type 4 changes, with a mean of 1,percentage points.
apparent retentions consistently overestimated true reten- However, when retentions were calculated in meat plus
tions, and these overestimates were frequently very large. skin of the turkey carcass, rather than in meat alone, dif-
ferent results were obtained (Table VII). In the same way
TYPE 5 CHANGE, SOLIDS AND MOISTURE LOST as calculations on meat only were made, calculation proce-
FROM MORE THAN ONE TISSUE dures allowed for proportions of light meat to dark meat to
Data were available for calculating retentions of proxi- skin, as determined by weights of these tissues in the car-
mate components, B-vitamins, cholesterol, and minerals in cass. For all 16 nutrients, differences between the two cal-
carcasses of turkeys of nine different age-sex groups. For 6 culation procedures were significant ( P = 0.05). The differ-
of the 41 replications, carcasses were separated into the ences ranged from 3 to 17 percentage points, with a mean
major parts prior to roasting. Parts from one side of each of 6 points, and in every case the apparent retention was
bird were reserved for analysis in the raw state, and parts higher.
from the other side were roasted. For the remaining 35 rep-
lications, carcasses were split in half, with one half being TYPE 6 CHANGE, SOLIDS GAINED AND MOISTURE
analyzed raw and the opposite half being analyzed after LOST
roasting. Each of the first six replications consisted of ten No data were available for calculating retentions of nu-
half-carcasses, and each of the remaining 35 replications trients in foods undergoing type 6 changes, such as dough-
included four half-carcasses. Roasting was done in alumi- nuts or french-fried potatoes, which take up fat while los-
num pans in ovens set a t 145 "C (325 "F) until the temper- ing moisture.
ature of the meat reached 85 "C (185 O F ) . Drippings were
weighed and saved for analysis. Analyses showed that DISCUSSION
moisture, fat, protein, and ash were all lost to the drip- For a number of nutrients posted in Tables I11 through
pings. VII, retentions appeared to be unusually high or low. For
Comparisons of apparent and true retentions were made instance, the retention of ash in oven-roasted peanuts,
for meat and for meat plus skin in the turkey carcasses on Table 111, was high, and the retention of sodium was low.
the same 16 nutrients as were determined for turkey gi- Derise et al. (1974) have suggested that the low sodium
blets. Table VI1 compares apparent and true retentions for value might be explained by loss of sodium into the peanut
16 nutrients in the meat only from the turkey carcasses. shells and hulls with heating. The high ash retentions were
Both calculation procedures have allowed for proportions not explained; possibly they represent problems in method-
of light meat to dark meat as determined by weights of ology of determining ash in raw compared with roasted
these tissues in the carcass. As can be seen in the table, peanuts. Crude fiber retentions shown in both Tables IV
there was little difference between results from the two cal- and V were also high. Data for fiber were obtained from
culation procedures for meat only. Of the 16 nutrients de- two widely separated laboratories which were not in com-
termined, only fat showed significant differences between munication with each other. Therefore, if the high fiber re-
the two calculation methods, and these differences were tentions indicate inaccurate methodology, the difficulty is
numerically small. For all 16 nutrients, the differences be- likely to be a general problem in applying the method for
J. Agric. Food Chem., Vol. 23,No. 6, 1975 1155
MURPHY, CRINER, GRAY
Table VII. Apparent and T r u e Retentions of Selected Nutrients in Roasted Turkey Carcasses
(Type 5a Weight Change with Cooking)
Meat only Meat plus skin
Apparent retention True retentionb Apparent retention True retentionb
No. of
Nutrient samples Mean,% C.V/ Mean,% C.V. Mean,% C.V. Mean,% C.V.
Protein 41 99 0.4 101 0.6 105 0.6 101 0.5
Fat 41 128 1.8 130 2.4 94 1.4 90 1.8
Ash 41 81 1.o 82 1.1 87 0.8 84 1.o
Thiamine 41 68 3.9 68 3.6 71 3.2 68 3.5
Ribof 1avine 41 82 2.7 83 2.6 90 2.6 83 3.4
Niacin 41 89 2 .o 90 2.2 96 2.2 92 2.3
Cholesterol 41 89 2.3 90 2.2 92 1.9 88 2.1
Calcium 41 130 2.8 132 2.8 137 2.5 13 1 2.6
Copper 9 71 12.8 71 12.2 84 11.3 72 11.8
Iron 41 96 3.8 97 3.9 101 3.9 97 4 .O
Magnesium 41 79 0.9 80 1.o 87 0.9 83 0.9
Manganese 9 83 8.7 84 8.4 102 7.4 87 8.1
Phosphorus 41 81 1.2 82 1.2 88 1.o 84 1.2
Pot as s ium 41 75 1.o 76 1.1 81 0.9 77 1.o
Sodium 41 76 1.2 77 1.2 82 1.1 79 1.2
Zinc 9 100 3.5 101 3.1 118 3 .O 101 2.8
Solids and moisture lost from more than one tissue. For meat only, differences between apparent and true retentions significant only
for fat ( P = 0.05); for meat plus skin, all differences significant ( P = 0.05). c Coefficient of variation,
crude fiber to both raw and cooked foods, rather than im- variation were 5.3% for true retentions and 4.9% for appar-
proper adaptation of a satisfactory method by a particular ent retentions. Furthermore, as was shown in several of the
laboratory. High fat and calcium retentions in Table VI1 previous examples, few individual comparisons of coeffi-
could possibly be caused by cooking of nutrients out of skin cients of variation showed appreciable differences between
or bone into the meat. Low retentions in Tables V, VI, and the two methods of calculation.
VI1 could be attributed to loss of nutrients to cooking water With the data a t hand, it is not possible to predict either
or drippings, or, for labile nutrients such as thiamin and the significance or the magnitude of differences between
ascorbic acid, to inactivation by heat. Regardless of wheth- true and apparent retentions. Therefore, it is not now pos-
er or not the retention data indicated problems in method- sible to establish correction factors which could be applied
ology, comparisons of apparent and true retentions, which to apparent retentions so that they would more closely esti-
were calculated on the same sample, are valid, and conclu- mate true retentions. Because of the improvement in accu-
sions about the relative accuracy of the two methods of cal- racy, with little change in variability of the data, true re-
culation can be drawn from the data reported here. tentions are preferable to apparent retentions in evaluating
Data in Tables 1 , 111, and IV indicate that for foods
1 the effects of cooking. The improved data to be obtained by
undergoing type 1 or type 2 changes (loss or gain of mois- using true retentions would be well worth the additional ef-
ture), apparent retentions, calculated on the dry basis, may fort required to weigh the raw and cooked foods. Of course,
or may not be significantly different from true retentions, in cases where it is not feasible to obtain batch weights be-
which take into account weight changes with cooking. Re- fore and after processing, such as on canning lines, true re-
tention data evaluated for foods exhibiting more complex tentions cannot be calculated, and some other approach to
changes with cooking showed that apparent retentions their estimation needs to be developed.
were not reliable estimates of true retentions. Apparent re- These findings, that apparent retentions in many in-
tentions, which tended to give false high values, did not stances are not reliable estimates of true retentions, are not
allow for loss of solids. Even for turkey carcass meat, the new. Streightoff et al. (1946), Dodds et al. (1944, 1946), and
one example of a complex cooking change in which the two Hewston et al. (1948) all judged that measurement of true
calculation procedures agreed well, weighting of the differ- retentions requires information on weight changes with
ent tissues in the food was necessary to arrive a t retentions cooking, if solids are lost with cooking or ingredients are
reasonably representative of all of the edible part of the added to the raw food. Watt and Attaya (1945), in a review
carcass. If weights of tissues are available, it would seem of published and unpublished data then available on reten-
reasonable to calculate true retentions rather than appar- tions of vitamins in quantity cooking of vegetables, also
ent retentions, which may be less accurate. warned that inaccuracies could result from using retentions
Even in those instances in which differences between the calculated on the dry basis. Few present-day food scientists
two calculation methods were not significant, apparent re- are familiar with the older literature on experimental cook-
tentions were almost always higher than true retentions. ery, including that on retentions of nutrients, except
The use of apparent retentions thus introduces a source of through such reviews as Harris and Von Loesecke’s “Nutri-
bias which could be eliminated by the use of true reten- tional Evaluation of Food Processing” (1960). Unfortunate-
tions. ly, this publication contains no discussion of problems in
The usefulness of a calculation system can be affected by calculating retentions.
the amount of variability it allows, in addition to its accu- Many research reports do not explain how retentions
racy. Coefficients of variation for apparent and true reten- were calculated, nor is it always possible, with the data
tions were therefore reviewed to see whether or not they given, for the reader to know what procedures were used.
were within reasonable bounds and if they differed appre- The findings reported here, which are based on new data,
ciably. For 115 comparisons, the average coefficients of confirm the judgments of the earlier researchers. It is
1156 J. Agric. Food Chem., Vol. 23,No. 6, 1975
BIOTIN CONTENT OF FEEDSTUFFS
hoped that this paper will encourage food scientists to Derise, N. L., Lau, H. A,, Ritchey, S. J., Murphy, E. W., J. Food
make their data more accurate and meaningful to others by Sci, 39,264-266 (1974).
Dodds, M. L., MacLeod, F. L., Carr, J. S., Tennessee A ricultural
reporting procedures used for calculating retentions as part Experiment Station, Progress Note 2, University of #ennessee,
of the description of analytical methods. True retentions, Knoxville, Tenn., June 1946.
rather than apparent retentions, should be reported when- Dodds, M. L.,MacLeod, F. L., Smith, J., Tennessee A ricultural
ever possible. Experiment Station Progress Note 1, University of +ennessee,
Knoxville, Tenn., Oct 1944.
ACKNOWLEDGMENT Fiske, C. H., Subbarow, Y., J . Bid. Chem. 66(2),375-400 (1925).
Freed, M., “Methods of Vitamin Assay”, 3rd ed, Interscience, New
Apparent and true retentions reported in this paper were York, N.Y., 1966.
calculated from results of laboratory research made under Harris, R. S., Von Loesecke, H., “Nutritional Evaluation of Food
the sponsorship of the Agricultural Research Service and Processing”, Avi Publishing Co., Inc., Westport, Conn., 1960,re-
printed 1971.
conducted a t the following locations: the University of Ha- Hewston, E. M., Dawson, E. H., Alexander, L. M., Orent-Keiles,
waii (R. L. Van Reen, Principal Investigator, and Mrs. Nao E.,.U.S. Dep. Agric.“MMisc. Publ. No. 628 (1948).
S. Wenkam), the University of Nebraska (T. E. Hartung, Perkin-Elmer Corp., Analytical Methods for Atomic Absorption
Spectrophotometry”, Norwalk, Conn., 1968.
Principal Investigator), and Virginia Polytechnic Institute Streightoff, F., Munsell, H. E., Ben-Dor, B. A., Orr, M. L., Cail-
and State University (S. Ritchey, Principal Investigator).
J. leau, R., Leonard, M. H., Ezekiel, S. R., Kornblum, R., Koch, F.
G., J. Am. Diet. Assoc. 22(2),117-127 (1946).
LITERATURE CITED T u , C., Powrie, W. D., Fennema, O., J . Food Sci. 32(1), 30-34
Ames, S. R., Risley, H. A., Harris, P. L., Anal. Chem 26,1378-1381 (1967).
(1954). Watt, B. K., Attaya, M. B., J . Home Econ. 37(6),340-344 (1945).
Association of Official Analytical Chemists, “Official Methods of
Analysis”, 11th ed, Washington, D.C., 1970. Received for review March 21,1975.Accepted July 23,1975.
Biotin Content of Feedstuffs
Jacob Scheiner* and Elmer De Ritter
The biotin contents of a’ variety of feedstuffs are 2 N acid for plant materials and with 6 N acid for
reported. Preliminary experiments using hydroly- samples of animal origin. Peanut meal, safflower
sis for 2 hr a t 121’ with 2 N and 6 N H&04 indi- seed meal, streptomyces meal and solubles, brew-
cated that higher results were obtained with 2 N ers’ yeast, dried liver, and a whey-yeast product
acid for feedstuffs of plant origin and with 6 N for had relatively high biotin contents. Other samples
feedstuffs ‘of animal origin. On the basis of these have been grouped in order of decreasing biotin
results, all subsequent extractions were made with contents.
A number of investigators (Patrick e t al., 1942; McGinnis assay using Lactobacillus plantarum (arabinosus 17-5,
and Carver, 1947; Roblee and Clandinin, 1953; Slinger and ATCC no. 80141, the test organism considered to yield the
Pepper, 1954) have reported biotin deficiency in poults fed most reliable results.
rations containing practical feed ingredients. However, the For the preparation of extracts for microbiological assay,
occurrence of this deficiency in commercial flocks was ei- no single hydrolytic procedure is universally effective for
ther not recognized or not reported until recently. It had maximum liberation of bound biotin. Table I summarizes
been generally believed that the feedstuffs in use, com- the results of various acid extraction procedures employed
bined with biotin arising from intestinal synthesis, sup- by a number of investigators. These studies indicate that
plied sufficient biotin to meet the poults’ requirement. Re- stronger acid concentrations are required to liberate bound
cently, however, the occurrence of biotin deficiency in com- biotin from animal tissues than from plant tissues. In the
mercial flocks has been reported (Brown, 1966; Wilson, extraction of plant tissues, biotin is less stable in relation to
1967; Richardson and Wilgus, 1967; Johnson, 1967). Mar- autoclaving time and acid concentration than in extraction
usich et al. (1970) encountered biotin deficiency symptoms from animal tissues.
in poults fed a commercial ration in the laboratory. Appar-
ent biotin deficiencies in swine under commercial condi- METHODS
tions have also been reported (Adams et al., 1967; Cunha et The microbiological assay procedure for biotin was that
al., 1968). As a consequence of these findings, a reevalua- of Wright and Skeggs (1944) with the exception that the
tion of the biotin content of feedstuffs was desirable, par- test organism was grown on the liver-tryptone agar of
ticularly since the available published data cover only a Nymon and Gortner (1946). Inocula were prepared from
limited number of feedstuffs and some of the results were stab cultures transferred the previous day,
obtained by methods whose validity could be questioned. In view of the demonstrated effects of acid concentration
The present study was undertaken to provide more com- and conditions of hydrolysis on yields of biotin from differ-
prehensive data on the biotin content of a variety of feed- ent materials, two hydrolytic procedures were used in the
stuffs. Biotin determinations were made by microbiological present study, namely, autoclaving for 2 hr a t 121’ with ei-
ther 2 N or 6 N HzS04. In each case, 20 ml of acid was used
per gram of sample. Similar conditions were used for ex-
traction of a number of samples with water as a means of
Hoffmann-La Roche Inc., Nutley, New Jersey 07110. estimating the content of free biotin. Recovery tests were
J. Agric. Food Chem., Vol. 23,No. 6, 1975 1157